I'm not "asking which is better". I realize there are often "un-measurables" and wonder if anyone had isolated aspects that may account for this ... Time alignment, freq. response(as stated earlier. Some like overstated bass, etc ...)
Re: "ears are best" - I take issue when I consider much of what people listen to these days.
Re: "ears are best" - I take issue when I consider much of what people listen to these days.
test microphone setup
Nobody can deny that
Linkwitz is one authorithy in LoudSpeakers.
We use his name to refer to the Linkwitz-Riley speaker x-over filter.
Though the years he has produced numerous papers & investigations on the subject.
some stuff at his personal website.
A dozen ways to evaluate a loudspeaker.
http://www.linkwitzlab.com/Loudspeaker evaluation.htm
PHOENIX.
Build your own open-baffle loudspeaker system.
System Test.
Including test microphone construction and shaped toneburst testing.
http://www.linkwitzlab.com/sys_test.htm
By www.diyaudio.com member .....
Here is the correct way to do speaker testing
Photo by fotios our man in Greece (Hellas)
Nobody can deny that
Linkwitz is one authorithy in LoudSpeakers.
We use his name to refer to the Linkwitz-Riley speaker x-over filter.
Though the years he has produced numerous papers & investigations on the subject.
some stuff at his personal website.
A dozen ways to evaluate a loudspeaker.
http://www.linkwitzlab.com/Loudspeaker evaluation.htm
PHOENIX.
Build your own open-baffle loudspeaker system.
System Test.
Including test microphone construction and shaped toneburst testing.
http://www.linkwitzlab.com/sys_test.htm
By www.diyaudio.com member .....
Here is the correct way to do speaker testing
Photo by fotios our man in Greece (Hellas)
Attachments
I have this kind of dilemma, every time when I work at a new loudspeaker design.
First of all, it is vey important to know what it really happen.
Hearing it is a little bit tricky, and it isn't the best way to start a project, in my opinion. It is true, we make loudspeakers for music, but when we make something like that, we need more than a "golden ear" .
Almost everybody run for a linear response but I am not sure if it is the best way to design a good loudspeaker. For me, a non linear loudspeaker sound better than ultra linear cabinets. Sometime, for classic music, a linear response sound better, but for pop music or jazz, a little deep ( - 1,5dB to 3dB ) will be a very good idea.
Anyway, to be on topic, the measurement don't exclude other ways to design a cabinet. I know few people how made very good loudspeakers, without any kind of measurement.
First of all, it is vey important to know what it really happen.
Hearing it is a little bit tricky, and it isn't the best way to start a project, in my opinion. It is true, we make loudspeakers for music, but when we make something like that, we need more than a "golden ear" .
Almost everybody run for a linear response but I am not sure if it is the best way to design a good loudspeaker. For me, a non linear loudspeaker sound better than ultra linear cabinets. Sometime, for classic music, a linear response sound better, but for pop music or jazz, a little deep ( - 1,5dB to 3dB ) will be a very good idea.
Anyway, to be on topic, the measurement don't exclude other ways to design a cabinet. I know few people how made very good loudspeakers, without any kind of measurement.
HiFiNutNut said:
Excellent thought, but I will add that while you may trust your ears, can you really trust your mind. There are tremendous psycho-acoustic factors that come into play when listening to a speaker or pair of speakers.
You would assume if you had a reference standard speaker, and compared all speakers against it for sound, then you would be OK, but your mind will always prefer the speaker that is slightly louder.
Psychologically, people will immediately prefer a speaker that has heavier bass. Though upon extended listening they may discover that it is an exaggerated one-note bass.
Some times a speaker with a tweeter that is a little louder, will jump out and be presumed to sound better.
Even people who listen to and review hundreds of speakers every year, still have to struggle to rein in the instinctive psycho-acoustic effects that can bias them with regard to a speaker.
I concede that some speakers that test right, don't sound right. And some speaker that sound right, don't test right. Consequently I think the truest test is BOTH measure and listen. Measuring give you some assurance that nothing is out of order. Listening verifies that a theoretical or technical idea was able to be translated into a working practical idea.
Just some random thoughts.
Steve/bluewizard
Yes, sort of "been there, done that".
For the same speaker, I could find that at one time it sounds perfect, then the next day awful.
When tuning a XO, I would keep trimming down the tweeter until I thought the harshness and brightness were completely gone, then the next day found them sounding terribly dull. So the tweeter was tuned up and up again, until the speakers sounded really "phenomenal" with a lot of presence and I was very happy. Then the next day I found harshness again. This cycle was repeated many times, sometimes tunning the bass, sometimes the treble, sometimes the BBC dip, etc, etc.
But still, eventually, experience can be gained during the repeated processes. I listen to live classical music and that serves as the reference. I play music CDs of string quartet and walk around the rooms and also listen to the music from the next room and ask myself: "does it sound like a real quartet play in the next room?" if it doesn't, something still needs to be done. The recordings could be at fault so multiple good recordings must be used.
I found it very educational to listen to my familiar music while adjusting an equalizer (you can do it from Wimamp from a PC with a good earphone) and understand the effects of the dips and peakers of frequency response and different sets of response profiles. This helps identify the peaks and dips of my speakers.
It is extremely tedious to do it this way and it took me a few years to tune a pair of speakers to near perfection.
It could have been substantially easier if the tunning was based on accurate measurement.
Regards,
Bill
For the same speaker, I could find that at one time it sounds perfect, then the next day awful.
When tuning a XO, I would keep trimming down the tweeter until I thought the harshness and brightness were completely gone, then the next day found them sounding terribly dull. So the tweeter was tuned up and up again, until the speakers sounded really "phenomenal" with a lot of presence and I was very happy. Then the next day I found harshness again. This cycle was repeated many times, sometimes tunning the bass, sometimes the treble, sometimes the BBC dip, etc, etc.
But still, eventually, experience can be gained during the repeated processes. I listen to live classical music and that serves as the reference. I play music CDs of string quartet and walk around the rooms and also listen to the music from the next room and ask myself: "does it sound like a real quartet play in the next room?" if it doesn't, something still needs to be done. The recordings could be at fault so multiple good recordings must be used.
I found it very educational to listen to my familiar music while adjusting an equalizer (you can do it from Wimamp from a PC with a good earphone) and understand the effects of the dips and peakers of frequency response and different sets of response profiles. This helps identify the peaks and dips of my speakers.
It is extremely tedious to do it this way and it took me a few years to tune a pair of speakers to near perfection.
It could have been substantially easier if the tunning was based on accurate measurement.
Regards,
Bill
To give my 2 cents to the original question.
It depends on if you want "good / audiophile" type of sound or "pure / original" type of sound.
If you are into "good" sound, then as small as 3dB up to 8dB gradual roll-off from 100Hz to 20kHz would do a lot of good. The sound would appear to have a lot more "texture, content and richness" and less brightness and harshness.
A dip of 1dB to 3dB around the 2k-4k region would make the music having more depth and more 3D, trading off "presence".
If the response curve is not flat, then the "audiophile" curve would be, starting from low freqencies ending up in high frequencies, relatively, 100-350Hz Up, 350Hz-750Hz down, 750Hz-1kHz up, 1kHz-4kHz down, 4kHz -8kHz up, 8kHz-12kHz up, 12kHz-20kHz down. I would not make the ups and downs more than 1dB or 2, however, the curve should be gradually rolling off from 100Hz to 20kHz for a least 3dB.
This would make many types of contempory music sould more "musical". However, if you go too far (more than 1dB or 2dB broad band dips and peaks), classical music may no longer be convincing. For classical music, you could introduce a "flat" roll off up to 6dB, but broad band deviation of even as small as 0.5dB or 1dB could sometimes make the music less emotionally envolving. Sharp nulls (high Q dips) are generally not audible. All peaks, low or high Q, are very audible.
Regards,
Bill
It depends on if you want "good / audiophile" type of sound or "pure / original" type of sound.
If you are into "good" sound, then as small as 3dB up to 8dB gradual roll-off from 100Hz to 20kHz would do a lot of good. The sound would appear to have a lot more "texture, content and richness" and less brightness and harshness.
A dip of 1dB to 3dB around the 2k-4k region would make the music having more depth and more 3D, trading off "presence".
If the response curve is not flat, then the "audiophile" curve would be, starting from low freqencies ending up in high frequencies, relatively, 100-350Hz Up, 350Hz-750Hz down, 750Hz-1kHz up, 1kHz-4kHz down, 4kHz -8kHz up, 8kHz-12kHz up, 12kHz-20kHz down. I would not make the ups and downs more than 1dB or 2, however, the curve should be gradually rolling off from 100Hz to 20kHz for a least 3dB.
This would make many types of contempory music sould more "musical". However, if you go too far (more than 1dB or 2dB broad band dips and peaks), classical music may no longer be convincing. For classical music, you could introduce a "flat" roll off up to 6dB, but broad band deviation of even as small as 0.5dB or 1dB could sometimes make the music less emotionally envolving. Sharp nulls (high Q dips) are generally not audible. All peaks, low or high Q, are very audible.
Regards,
Bill
At the AVGuide forums there is a specific subject thread on this topic worth browsing. Some very good posts there. Within the post from Skeptic67 there is a link to a Dr. Floyd Toole of JBL on subjective vs objective studies he's conducted.
To read more about it....
http://forums.avguide.com/viewforum.php?f=144&sid=09fedc8d9ccf6b1e791d4700bbf03941
To read more about it....
http://forums.avguide.com/viewforum.php?f=144&sid=09fedc8d9ccf6b1e791d4700bbf03941
I've built some speakers that were ruler flat on axis, but very unsatisfying. I can't imagine building a speaker without both measurement and listening tests, and over quite a long time. No good speaker was created in a week. I also voice a speaker differently for CD vs vinyl. Never found a speaker that was optimal for both.
HiFiNutNut said:
This is pretty much my experience too.
When you say, "...then the next day found them sounding terribly dull." I think what is happening is that when you listen you are always comparing what you hear now with a memory of what they sounded like previously, or an idea of what they should sound like. In either case what is in your head is not an absolute, and so your judgement cannot be absolute either.
If you repeat the process you may get closer to the truth but you may go around in circles, it is a problem where measurements can help as you have something absolute to reference against.
So I think listen and measure.
I was at a speaker demonstration this evening, and it left me a little confused. The main snafu of the evening was having a sales man give a demonstration to a room full of loudspeaker designers and recording engineers - the presentation fell pretty flat.
But it made me feel at a loss about how to work with audio, particularly in terms setting goals and voicing a final product. The audience tonight really wanted to talk details (ferrofluid, definition of compression, etc) while the marketing guy just wanted to impress us with movie soundtracks and good vibes. One one hand, I felt like focusing on the details misses the overall picture of enjoying music, and the practical limits of doing so, while playing fancy soundtracks awash in compression and reverb leaves you totally at sea when it comes to any sort of design practicality. I suppose the most mature designers would have an exceptional grasp of the intricacies of acoustic transduction and sensual appreciation.
I guess nowadays I find strong opinions from one type as equally obnoxious as from the other...
WRT the OP's questions, there seems to be no consensus among people about how to design or test anything in audio. I can't think of any field so fraught with internal conflict. The best I can suggest is to educate yourself as much as possible and do whatever you enjoy. Thats what I try at least...
But it made me feel at a loss about how to work with audio, particularly in terms setting goals and voicing a final product. The audience tonight really wanted to talk details (ferrofluid, definition of compression, etc) while the marketing guy just wanted to impress us with movie soundtracks and good vibes. One one hand, I felt like focusing on the details misses the overall picture of enjoying music, and the practical limits of doing so, while playing fancy soundtracks awash in compression and reverb leaves you totally at sea when it comes to any sort of design practicality. I suppose the most mature designers would have an exceptional grasp of the intricacies of acoustic transduction and sensual appreciation.
I guess nowadays I find strong opinions from one type as equally obnoxious as from the other...
WRT the OP's questions, there seems to be no consensus among people about how to design or test anything in audio. I can't think of any field so fraught with internal conflict. The best I can suggest is to educate yourself as much as possible and do whatever you enjoy. Thats what I try at least...
Though slightly tangential to the original subject, I think we need to remember that the subject is far far more complex than good and not good.
I'm always on a strict budget, so when I have speaker related questions, and a guy who typically builds and listens to four figure speakers, tells me what I should and shouldn't do, I really can't give it that much weight.
Why? Because my $100 budget question can not be satisfied with $1000 solutions.
So, in a sense, a speaker is never good, at best it is good for the money. If you spend $1000 and it sounds like $5000, then that is a good speaker. But if the opposite is true, spend $5000 and it sounds like $1000, no matter how good the speaker sounds, it's not a good speaker for the money.
My point is that 'good' isn't black and white, at best it is many many shades of gray.
So, my 'good' speaker could very easily be your 'bad' speaker, and your 'good' speaker could be my 'bad' speaker (though mostly because the cost would bankrupt me).
And sometimes people actually like bad speakers. So buy any standard, a generally bad speaker is actually their good speaker.
A fair, though not perfect, illustration of this is crossovers. I say that for a beginner, it is not a sin to buy crossovers off the shelf. Though I am well aware that many here will consider that sacrilege. One step above that is 'cookbook' crossovers, but again many here will say they are worthless.
But right now take an assessment of the resources necessary to build a custom and very precisely tuned crossover. That is a huge outlay of resources in terms of time, equipment, education, and money. Something that most beginners simply don't have.
But as HiFiNutNut implies, building a speaker is a job that is never truly done. I've been building my current set of speakers since 1985 and they still aren't done. They keep getting better, but they still aren't done.
Our speakers can sound very very good, but on some level, they never sound good enough. We always feel there is a way to tweak a little more out of them.
That's why I don't hesitate to tell someone to start with off-the-shelf crossovers. If they do and are satisfied, then they have found their personal 'good for the money'. If not then I know a couple of months or years down the road off-the-shelf are going to be traded for cookbook', and if that is their 'good' then fine. And if not, they they are likely to start testing, which means accumulating equipment and knowledge in how to use the equipment. And that in turn is going to lead to a better understanding of speaker and how to build them.
And eventually they will reach the level that satisfies them. Keep in mind that one can spend $250,000 on a pair of speakers. That makes our $2500 speakers seem trivial, and our $250 speakers seem like K-Mart blue light specials.
So, regardless of what tests good, or what sounds good, you aren't going to stop until we find our own personal 'good' or run out of money trying.
Just another perspective.
Steve/bluewizard
I'm always on a strict budget, so when I have speaker related questions, and a guy who typically builds and listens to four figure speakers, tells me what I should and shouldn't do, I really can't give it that much weight.
Why? Because my $100 budget question can not be satisfied with $1000 solutions.
So, in a sense, a speaker is never good, at best it is good for the money. If you spend $1000 and it sounds like $5000, then that is a good speaker. But if the opposite is true, spend $5000 and it sounds like $1000, no matter how good the speaker sounds, it's not a good speaker for the money.
My point is that 'good' isn't black and white, at best it is many many shades of gray.
So, my 'good' speaker could very easily be your 'bad' speaker, and your 'good' speaker could be my 'bad' speaker (though mostly because the cost would bankrupt me).
And sometimes people actually like bad speakers. So buy any standard, a generally bad speaker is actually their good speaker.
A fair, though not perfect, illustration of this is crossovers. I say that for a beginner, it is not a sin to buy crossovers off the shelf. Though I am well aware that many here will consider that sacrilege. One step above that is 'cookbook' crossovers, but again many here will say they are worthless.
But right now take an assessment of the resources necessary to build a custom and very precisely tuned crossover. That is a huge outlay of resources in terms of time, equipment, education, and money. Something that most beginners simply don't have.
But as HiFiNutNut implies, building a speaker is a job that is never truly done. I've been building my current set of speakers since 1985 and they still aren't done. They keep getting better, but they still aren't done.
Our speakers can sound very very good, but on some level, they never sound good enough. We always feel there is a way to tweak a little more out of them.
That's why I don't hesitate to tell someone to start with off-the-shelf crossovers. If they do and are satisfied, then they have found their personal 'good for the money'. If not then I know a couple of months or years down the road off-the-shelf are going to be traded for cookbook', and if that is their 'good' then fine. And if not, they they are likely to start testing, which means accumulating equipment and knowledge in how to use the equipment. And that in turn is going to lead to a better understanding of speaker and how to build them.
And eventually they will reach the level that satisfies them. Keep in mind that one can spend $250,000 on a pair of speakers. That makes our $2500 speakers seem trivial, and our $250 speakers seem like K-Mart blue light specials.
So, regardless of what tests good, or what sounds good, you aren't going to stop until we find our own personal 'good' or run out of money trying.
Just another perspective.
Steve/bluewizard
cuibono said:
or any aspect of musicality
gees I wish these sales rooms would get with the program !!recorded live acoustics, chamber, jazz... with minimal "manipulation"... much better for judging a speaker.
planet10 said:
Better yet, just forget the XO and build a 1-way
dave
Yep, then you don't really have to think about the top end.... or the bottom end for that matter !
Hi Dave !
Andy G said:
You might be surprised... the big thing is that the midrange isn't mucjed up by an XO... my latest thou is a small active (PLLXO) 2-way. XO ended up at 333 Hz. Haven't measured this iteration, but an earlier rev was -/- 2dB 25 hz to the limit of my measuring kit (16k, but someone elses verified that response was out past 20k). Really a stunning little beast.
dave
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.